Association of Preoperative Anemia and Perioperative Allogenic Red Blood Cell Transfusion with Oncologic Outcomes in Patients with Nonmetastatic Colorectal Cancer

ANEMIA AND TRANSFUSION IN COLORECTALORIGINAL CANCER, ARTICLE Kwon et al.

Association of preoperative anemia and perioperative allogenic red blood cell transfusion with oncologic outcomes in patients with nonmetastatic colorectal cancer

†‡ § H.Y. Kwon md,* B.R. Kim md, and Y.W. Kim md phd*

ABSTRACT

Background We investigated whether preoperative anemia and perioperative blood transfusion (pbt) are associated with overall survival and recurrence-free survival in patients with nonmetastatic colorectal cancer.

Methods From 1 January 2009 to 31 December 2014, 1003 patients with primary colorectal cancer were enrolled in the study. Perioperative clinical and oncologic outcomes were analyzed based on the presence of preoperative anemia and pbt.

Results Preoperative anemia was found in 468 patients (46.7%). In the anemia and no-anemia groups, pbt was performed in 44% and 15% of patients respectively. Independent predictors for pbt were preoperative anemia, higher American Society of Anesthesiologists score, laparotomy, lengthy operative time, advanced TNM stage, T4 stage, and 30-day morbidity. The use of pbt, but not preoperative anemia, was found to be an independent adverse prognostic factor for overall survival. In terms of recurrence-free survival, the presence of preoperative anemia was similarly not a significant prognostic factor, but the use of pbt was an independent factor for an unfavourable prognosis.

Conclusions The use of pbt, but not preoperative anemia, was independently associated with worse overall and recurrence-free survival in nonmetastatic colorectal cancer. For better oncologic outcomes, our findings indicate a need to reduce the use of blood transfusion during the perioperative period.

Key Words Colonic neoplasms, rectal neoplasms, anemia, blood transfusions, survival

Curr Oncol. 2019 June;26(3):e357-e366 www.current-oncology.com

INTRODUCTION homeostasis related to chronic systemic inflammation, which leads to anemia of chronic disease5,6. Studies of Colorectal cancer (crc) is the 3rd most common cancer in whether anemia in patients with crc is a negative prognos- men and the 2nd most common cancer in women1. Further- tic factor have produced conflicting results7–18. more, crc is the 4th most common cause of cancer-related Allogenic red blood cell (rbc) transfusion is a treatment death worldwide1,2. Anemia is a common condition in pa- for anemia that is frequently used in patients with crc. tients with crc, with an incidence of 38%–59% in patients However, concerns have been raised about the negative ef- with colon cancer and of 18%–50% in those with rectal fect of blood transfusion on oncologic outcomes. Immuno- cancer3. The World Health Organization defines anemia as modulation and systemic inflammatory responses related a hemoglobin level less than 13 g/dL in men and less than to blood transfusion are thought to be related to adverse 12 g/dL in women4. The main cause of anemia in patients oncologic outcomes19,20. Indeed, a number of studies have with crc is iron deficiency resulting from occult bleed- shown an adverse effect of blood transfusion on postoper- ing from tumours; another cause is impairment of iron ative outcomes in crc patients, including infection-related

Correspondence to: Young-Wan Kim, Department of Surgery, Yonsei University Won-ju College of Medicine, 20 Ilsan-ro, Won-ju-si, Gangwon-do, Won-ju 26426 R.O.K. E-mail: [email protected] n DOI: https://doi.org/10.3747/co.26.4983 Supplemental material available at http://www.current-oncology.com

Current Oncology, Vol. 26, No. 3, June 2019 © 2019 Multimed Inc. e357 ANEMIA AND TRANSFUSION IN COLORECTAL CANCER, Kwon et al. complications and survival21. Some investigators have for additional treatment within 30 days after surgery. suggested that perioperative blood transfusions do not The Clavien–Dindo classification system was used to influence overall or recurrence-free survival (rfs) after determine the severity of surgical complications26. A resection for crc17,22–24. condition requiring medical treatment was defined as a The association of preoperative anemia and periop- grade 2 complication. If surgical, endoscopic, or radiologic erative blood transfusion with oncologic outcomes is still intervention was required, the complication was defined disputed. In clinical practice, the influence of preoperative as a grade 3 complication (Clavien–Dindo classification anemia on outcomes is often confounded by periopera- system)27. Tumour recurrence was defined as any rative blood transfusion. To date, studies investigating the diologic or histologic evidence of tumour growth in the influence on crc outcomes of preoperative anemia with previous surgical field or distant organs. perioperative blood transfusion are lacking9,17. The aim of the present study was to investigate whether preoperative Study Objectives anemia and perioperative blood transfusion are associated The primary objective of the study was to investigate with overall survival (os) and rfs in patients undergoing whether preoperative anemia or perioperative allogenic curative resection for nonmetastatic crc. rbc transfusion is associated with oncologic outcomes (os and rfs) after curative surgery in patients with non- METHODS metastatic crc.

Patients Preoperative Chemoradiation, Surgery, Adjuvant Our study enrolled 1003 patients with histology-confirmed Therapy, and Follow-up primary crc who underwent elective crc surgery from Patients with clinical stage T3 and T4 or node-positive 1 January 2009 to 31 December 2014. This retrospective (or both) mid- or low rectal cancer received preoperative observational study, performed at a tertiary university chemoradiation28. A total dose of 50.4 Gy was delivered hospital, was approved by the institutional review board over 5 weeks. All surgeries for crc were performed by (ywmr-15-5-050). All clinical data were gathered in accor- colorectal surgery specialists29. The type of surgical ap- dance with the principles of the Helsinki Declaration. All proach (laparotomy, laparoscopy, or robot-assisted) was the work complied with the strobe (Strengthening the discussed preoperatively with patients and their families30. Reporting of Observational Studies in Epidemiology) guide- Complete mesocolic excision and central vascular ligation lines25. Eligibility criteria included histology-confirmed were performed for colon cancer; high ligation of the infe- crc, elective surgery, curative surgery for crc, and age older rior mesenteric artery and total mesorectal excision were than 20 years. Patients with distant organ metastases or performed for rectal cancer28,31. Adjuvant chemotherapy recurrent cancer, or those undergoing emergency surgery after recovery from surgery was recommended for patients or palliative non-resection surgery were excluded from with pathologic stage ii or iii crc based on guidelines from the study cohort. the U.S. National Comprehensive Cancer Network32,33. Clinical, laboratory, and pathology data, including Patients in the study cohort were followed until death or age, sex, American Society of Anesthesiologists (asa) clas- 30 June 2015. The median follow-up period was 41 months sification, TNM stage as defined by the 8th edition of the (interquartile range: 23.8–70.8 months). American Joint Committee on Cancer staging manual, tumour location, preoperative laboratory data, and use of Statistical Analysis perioperative transfusion were obtained from electronic All statistical analyses were performed using the IBM medical records and a dedicated institutional crc database. SPSS Statistics (version 23.0: IBM, Armonk, NY, U.S.A.) and MedCalc (version 17.4: MedCalc Software, Ostend, Belgium) Variables and Outcome Measures software applications. Categorical variables are presented as frequencies and percentages and were compared using Preoperative Anemia the chi-square test or Fisher exact test. Continuous vari- Based on the World Health Organization classification, ables are presented as means with standard deviation and anemia was defined as a hemoglobin level less than were analyzed using the Student t-test. 13 g/dL in male patients and less than 12 g/dL in female To determine the predictors of perioperative blood patients on preoperative complete blood counts4. transfusion, univariate logistic regression analyses were performed for all variables. Variables with a p value less Perioperative Allogenic RBC Transfusion than 0.05 (preoperative anemia, age, asa score, surgical Use of perioperative allogenic rbc transfusion was defined approach, operative time, TNM stage, depth of tumour as receipt at least 1 unit of packed rbcs during a patient’s invasion, and 30-day postoperative morbidity) were then in-hospital admission (from the time of primary surgery used in multivariable analyses. Multivariable logistic re- to hospital discharge). gression analyses were performed using forward stepwise selection of variables. Variables Survival curves were constructed according to the The right-side colon was considered to extend from Kaplan–Meier method and were compared using the log- the cecum to the transverse colon. The left-side colon rank test. The definition of os was the number of months considered to include the splenic flexure to the sigmoid from surgery to death, and the definition of rfs was the colon. Postoperative morbidity was defined as the need number of months from surgery to recurrence.

We used a Cox proportional hazards model to inves- p = 0.002), but not in multivariate analysis (hr: 1.06; 95% ci: tigate factors prognostic for os. The univariate Cox pro- 0.77 to 1.45; p = 0.729). The use of perioperative allogenic rbc portional hazards model analysis used all variables, and transfusion was an independent adverse prognostic factor variables with a p value less than 0.05 (preoperative anemia, (hr: 1.55; 95% ci: 1.12 to 2.13; p = 0.008). Other adverse prog- perioperative transfusion, age, asa score, surgical approach, nostic factors for os (Table iii) were laparotomy rather than TNM stage, 30-day postoperative morbidity, and adjuvant minimally invasive surgery (hr: 1.75; 95% ci: 1.26 to 2.41; p = chemotherapy) were entered into the multivariable analysis. 0.001); advanced TNM stage, including stage ii (hr: 2.99; 95% Likewise, Cox proportional hazards models were used to ci: 1.71 to 5.21; p < 0.001) and stage iii (hr: 6.12; 95% ci: 3.62 investigate factors prognostic for rfs. After the univariate to 10.32; p < 0.001); 30-day postoperative morbidity (hr: 1.51; analysis, variables with a p value less than 0.05 (preoperative 95% ci: 1.11 to 2.05; p = 0.009); and no adjuvant chemotherapy anemia, perioperative transfusion, age, surgical approach, (hr: 3.85; 95% ci: 2.66 to 5.59; p < 0.001). TNM stage, histology, and 30-day postoperative morbidity) were used in the multivariable analysis. The multivariable Cox Proportional Hazards Model of Factors Cox proportional hazards model analyses used forward Prognostic for RFS stepwise selection of variables. The presence of preoperative anemia was not a significant prognostic factor in either the univariate (p = 0.06) RESULTS or the multivariate analysis (p = 0.902). However, the use of perioperative allogenic rbc transfusion (hr: 1.65; 95% Patient Characteristics by Presence of ci: 1.24 to 2.20; p = 0.001), TNM stage iii (hr: 3.2; 95% ci: Preoperative Anemia 2.12 to 4.83; p < 0.001), and histologic grade 3 (hr: 1.8; 95% Of the 1003 study patients, 468 (46.7%) presented with ane- ci: 1.02 to 3.17; p = 0.041) were independent unfavourable mia. Of the patients in the anemia and no-anemia groups, prognostic factors for rfs (Table iv). 44% and 15% respectively received a perioperative allogenic rbc transfusion (p < 0.001). Compared with patients in the Kaplan–Meier Survival Analysis for Three-Year no-anemia group, patients in the anemia group were older Rates of OS and RFS (69.4 ± 11 years vs. 65.6 ± 11 years, p < 0.001), had a lower body The 3-year os rate was significantly worse in the anemia mass index (22.6 ± 3 vs. 24.1 ± 3, p < 0.001), had higher asa group than in the no-anemia group [81.5% vs. 87.5%, p = scores (3 or 4: 20% vs. 14%, p = 0.029), more frequently had 0.002, Figure 1(A)]. Compared with patients who did not right-side colon cancer (31% vs. 18%, p < 0.001), had higher receive perioperative allogenic rbc transfusions, those levels of serum carcinoembryonic antigen (8.7 ± 25 ng/mL vs. who received such transfusions had a worse 3-year os rate 5.1 ± 11 ng/mL, p = 0.005), had a more advanced TNM stage [76.9% vs. 88.2%, p < 0.001, Figure 1(B)]. When preoper- (stage iii: 47% vs. 39%, p < 0.001), and had tumours larger in ative anemia and perioperative transfusions were both size (5.4 ± 2 cm vs. 3.9 ± 2 cm, p < 0.001). More women (41% taken into consideration, the 3-year os rates were 76.1%, vs. 33%, p = 0.013) were included in the anemia group than 78.4%, 86.2%, and 89.4% in the anemia and transfusion, in the no-anemia group. In the anemia group, laparotomy the no-anemia and transfusion, the anemia and no-trans- was more frequent than minimally invasive surgeries such fusion, and the no-anemia and no-transfusion groups as laparoscopy and robot-assisted surgery (46% vs. 30%, respectively [p < 0.001, Figure 1(C)]. p < 0.001, Table i). The presence of preoperative anemia was not associated with a significantly worserfs rate [75.0% vs. 80.6%, p = 0.059, Predictors for Perioperative Allogenic Figure 2(A)]. However, receiving a perioperative allogenic RBC Transfusion rbc transfusion was associated with a significantly worse In the univariate analysis, preoperative anemia, age 80 years rfs rate [70.9% vs. 80.9%, p < 0.001, Figure 2(B)]. Overall, the or older, higher asa score (3 or 4), laparotomy rather than min- 3-year rfs rates were 71.3%, 76.7%, 78.6%, and 82.4% in the imally invasive surgery, longer operative time (≥300 minutes), no-anemia and transfusion, the anemia and transfusion, more advanced TNM stage, greater tumour invasion depth the anemia and no-transfusion, and the no-anemia and (T4), and 30-day postoperative morbidity were associated with no-transfusion groups respectively [p < 0.001, Figure 2(C)]. perioperative blood transfusion. In the multivariate analysis, preoperative anemia [hazard ratio (hr): 5.06; 95% confidence DISCUSSION interval (ci): 3.42 to 7.48; p < 0.001], higher asa score [3 or 4 (hr: 1.88; 95% ci: 1.19 to 2.98; p = 0.007)], laparotomy (hr: The major finding of the present study is that the perioper- 1.78; 95% ci: 1.20 to 2.63; p = 0.004), lengthy operative time ative use of blood transfusion is independently associated (hr: 3.74; 95% ci: 2.20 to 6.35; p < 0.001), advanced TNM stage with worse os and rfs after curative surgery in patients (p = 0.036), T4 stage (hr: 1.67; 95% ci: 1.01 to 2.76; p = 0.048), with nonmetastatic crc, but that preoperative anemia is and 30-day postoperative morbidity (hr: 2.86; 95% ci: 1.96 to not. Compared with patients not having anemia, those 4.19; p < 0.001) were independent predictors of perioperative with anemia had distinctive patient factors (older age, blood transfusion (Table ii). female sex, lower body mass index, higher likelihood of an asa score of 3, more predominant right-sided colon cancer, Cox Proportional Hazards Model of Factors more advanced TNM stage, higher level of carcinoembry- Prognostic for OS onic antigen, and larger tumour size) and treatment-related The presence of preoperative anemia was a significant factors (laparotomy). In our study, the use of perioperative factor in univariate analysis (hr: 1.58; 95% ci: 1.17 to 2.12; blood transfusion was independently associated with

Current Oncology, Vol. 26, No. 3, June 2019 © 2019 Multimed Inc. e359 ANEMIA AND TRANSFUSION IN COLORECTAL CANCER, Kwon et al. worse os and rfs, and the independent predictors for Although we could not identify a direct causative re- perioperative blood transfusion were inherent patient fac- lationship between perioperative transfusion and adverse tors (including preoperative anemia, greater asa score, ad- oncologic outcomes, we tried to show the actual effect of vanced TNM stage, and T4 disease) and treatment-related preoperative anemia and perioperative transfusion by con- factors (including laparotomy, lengthy operative time, and trolling for possible confounding variables in a multivariate 30-day postoperative morbidity). analysis. Our findings showed that perioperative blood

TABLE I Patient characteristics according to the presence of preoperative anemia

Characteristic Patients by anemia status p Value Overall Anemia No anemia

Patients (n) 1003 468 535 Perioperative allogenic RBC transfusion [n (%) yes] 295 (29) 207 (44) 88 (16) <0.001 Mean age (years) 67.3±11 69.4±11 65.6±11 <0.001 Sex [n (%) men] 630 (63) 275 (59) 355 (66) 0.013 Mean BMI (kg/m2) 23.4±3 22.6±3 24.1±3 <0.001 ASA score [n (%)] 1 195 (19) 79 (17) 116 (22) 0.029 2 635 (63) 294 (63) 341 (64) 3 168 (167) 91 (19) 77 (14) 4 5 (0) 4 (1) 1 (0) Preoperative chemoradiation [n (%)] 113 (11) 43 (9) 70 (13) 0.052 Tumour location [n (%)] Right-side colon 239 (24) 144 (31) 95 (18) <0.001 Left-side colon 270 (27) 121 (26) 149 (28) Rectum 469 (47) 186 (40) 283 (53) Multiple sites 25 (2) 17 (4) 8 (1) Mean serum CEA (ng/mL) 6.7±19 8.7±25 5.1±11 0.005 Surgical approach [n (%)] Laparotomy 378 (38) 215 (46) 163 (30) <0.001 Laparoscopy or robot 625 (62) 253 (54) 372 (70) Mean operative time (minutes) 212±94 206±105 217±83 0.118 TNM stage [n (%)] 0,I 253 (25) 79 (17) 174 (33) < 0.001 II 318 (32) 168 (36) 150 (28) III 432 (43) 221 (47) 211 (39) Histologic grade [n (%)] 1, 2 898 (90) 432 (92) 466 (87) 0.007 3, Other 105 (10) 36 (8) 69 (13) Retrieved lymph nodes (n) 21.8±12 24.1±12 19.8±11 <0.001 Mean tumour size (cm) 4.6±2 5.4±2 3.9±2 <0.001 30-Day postoperative ... Morbidity [n (%)] 309 (31) 146 (31) 163 (30) 0.803 Infectious morbidity [n (%)] 177 (18) 90 (19) 87 (16) 0.218 Mortality [n (%)] 11 (1) 4 (1) 7 (1) 0.491 Clavien–Dindo classification [n (%)] 1,2 173(56) 76 (52) 97 (59) 0.233 3–5 138(44) 70 (48) 68 (41) Adjuvant chemotherapy [n (%)] 607(61) 296 (63) 311 (58) 0.098

RBC = red blood cell; BMI = body mass index; ASA = American Society of Anesthesiologists; CEA = carcinoembryonic antigen.

e360 Current Oncology, Vol. 26, No. 3, June 2019 © 2019 Multimed Inc. ANEMIA AND TRANSFUSION IN COLORECTAL CANCER, Kwon et al. transfusion was an independent prognostic factor for os in which the development of obstructive symptoms is and rfs, but that preoperative anemia was not. Compared frequently delayed. Intraluminal bleeding from friable with findings in other studies, those results are unique. cancer mucosa might persist until tumour progression in patients without symptoms5,35. Considering that the ane- Preoperative Anemia mia in crc is explained by intraluminal tumour bleeding, In the present study, 46.7% of the patients presented with anemia is more likely to be a feature of advanced tumours, anemia. Anemia is more common in patients who are which are likely to be larger in size, associated with higher elderly and female17. Hemoglobin levels are affected by levels of carcinoembryonic antigen, and classified into a physiologic changes associated with aging, such as de- more advanced T stage16,17. Anemia of chronic disease is clining production of rbcs and a shortened rbc lifespan34. caused by a systemic inflammatory response tocrc 36 and Anemia occurs more frequently in right-side colon cancer, is frequently associated with sarcopenia, decreased body

TABLE II Logistic regression analysis for predictors of perioperative allogenic red blood cell transfusion

Variable Comparator Univariate Multivariate

HR 95% CI p Value HR 95% CI p Value

Preoperative anemia (yes) No 4.03 3.01 to 5.40 <0.001 5.06 3.42 to 7.48 <0.001 Age ≥80 years <80 Years 1.72 1.17 to 2.54 0.006 1.48 0.91 to 2.39 0.111 Male sex Female sex 0.90 0.68 to 1.19 0.448 ASA score 3 or 4 ASA score 1 or 2 1.88 1.34 to 2.64 <0.001 1.88 1.19 to 2.98 0.007 Tumour location (rectum) Colon 1.11 0.84 to 1.46 0.443 Surgical approach (laparotomy) Laparoscopy or robot 2.27 1.72 to 3.00 <0.001 1.78 1.20 to 2.63 0.004 Operative time ≥300 minutes <300 Minutes 2.64 1.71 to 4.10 <0.001 3.74 2.20 to 6.35 <0.001 TNM stage 0, I 0.001 0.036 II 2.04 1.40 to 2.97 <0.001 1.56 0.95 to 2.56 0.082 III 1.47 1.02 to 2.11 0.039 0.93 0.56 to 1.54 0.772 Depth of tumour invasion (T4) T1–3 1.62 1.15 to 2.29 0.006 1.67 1.01 to 2.76 0.048 30-Day postoperative morbidity (yes) No 2.59 1.95 to 3.45 <0.001 2.86 1.96 to 4.19 <0.001

HR = hazard ratio; CI = confidence interval; ASA, American Society of Anesthesiologists.

TABLE III Cox proportional hazards model for prognostic factors of overall survival

Variable Comparator Univariate Multivariate

HR 95% CI p Value HR 95% CI p Value

Preoperative anemia (yes) No 1.58 1.17 to 2.12 0.002 1.06 0.77 to 1.45 0.729 Perioperative transfusion (yes) No 1.91 1.42 to 2.57 <0.001 1.55 1.12 to 2.13 0.008 Age ≥80 years <80 Years 2.35 1.62 to 3.41 <0.001 1.36 0.90 to 2.05 0.148 Male sex Female sex 0.87 0.64 to 1.18 0.368 ASA score 3 or 4 ASA score 1 or 2 1.96 1.39 to 2.77 <0.001 1.16 0.81 to 1.67 0.422 Tumour location (right-side colon) 0.187 Left-side colon 0.97 0.62 to 1.52 0.885 Rectum 1.36 0.93 to 1.99 0.113 Multiple sites 1.45 0.57 to 3.69 0.441 Surgical approach (laparotomy) Laparoscopy or robot 2.22 1.64 to 3.02 <0.001 1.75 1.26 to 2.41 0.001 TNM stage 0, I <0.001 <0.001 II 1.46 0.88 to 2.42 0.141 2.99 1.71 to 5.21 <0.001 III 2.71 1.73 to 4.24 <0.001 6.12 3.62 to 10.32 <0.001 Histologic grade 3 or other Grade 1 or 2 0.60 0.32 to 1.11 0.104 30-Day postoperative morbidity (yes) No 1.98 1.48 to 2.65 <0.001 1.51 1.11 to 2.05 0.009 Adjuvant chemotherapy (yes) No 1.89 1.41 to 2.53 <0.001 3.85 2.66 to 5.59 <0.001

HR = hazard ratio; CI = confidence interval; ASA = American Society of Anesthesiologists.

TABLE IV Prognostic factors for recurrence-free survival using Cox proportional hazards model

Variable Comparator Univariate Multivariate

HR 95% CI p Value HR 95% CI p Value

Preoperative anemia (yes) No 1.28 0.99 to 1.66 0.06 0.98 0.74 to 1.30 0.902 Perioperative transfusion (yes) No 1.77 1.36 to 2.30 <0.001 1.65 1.24 to 2.20 0.001 Age ≥80 years <80 Years 1.48 1.02 to 2.13 0.037 1.45 1.00 to 2.09 0.05 Male sex Female sex 0.87 0.64 to 1.18 0.368 ASA score 3 or 4 ASA score 1 or 2 1.32 0.95 to 1.83 0.104 Tumour location (right-side colon) Left-side colon 1 0.67 to 1.49 0.998 Rectum 1.39 0.99 to 1.95 0.058 Multiple sites 2.07 1.01 to 4.23 0.046 Surgical approach (laparotomy) Laparoscopy or robot 1.54 1.19 to 2.00 0.001 1.3 0.99 to 1.70 0.057 TNM stage 0, I <0.001 <0.001 II 1.61 1.02 to 2.53 0.041 1.51 0.95 to 2.39 0.082 III 3.31 2.21 to 4.96 <0.001 3.2 2.12 to 4.83 <0.001 Histologic grade 3 or other Grade 1 or 2 1.92 1.1 to 3.36 0.023 1.8 1.02 to 3.17 0.041 30-Day postoperative morbidity (yes) No 1.46 1.12 to 1.91 0.005 1.24 0.94 to 1.62 0.123 Adjuvant chemotherapy (yes) No 1.06 0.81 to 1.39 0.659

HR = hazard ratio; CI = confidence interval; ASA = American Society of Anesthesiologists. mass index, and the presence of severe comorbidities37, as Several factors could account for those conflicting confirmed in our study. results. First, diverse definitions of “perioperative” have Studies of whether preoperative anemia leads to worse been used in the studies9,24,40,44–47; estimating the exact outcomes have produced conflicting results. Several reasons association between the use of blood transfusion and worse for those findings could be considered. First, the systemic oncologic outcomes is therefore difficult. Second, complex inflammatory response linked to anemia is associated with clinical circumstances necessitating blood transfusion tumour progression18. Anemia is associated with increased have been considered. It has been suggested that, not systemic inflammation, including a higher modified the transfusion itself, but the situation that necessitates Glasgow prognostic score and higher levels of C-reactive the blood transfusion is the real determinant of progno- protein and interleukin 87. Second, anemia is associated with sis23,24. But our study did not support that hypothesis. Our an increased occurrence of postoperative complications, results suggest that the use of perioperative transfusion is which are adversely associated with oncologic outcomes36. independently associated with worse oncologic outcomes Third, increased use of blood transfusions is associated regardless of the clinical circumstances. Third, there is with worse oncologic outcomes because of transfusion- a possibility that selection bias because of small study induced immunomodulation and systemic inflammatory samples or variation in study cohorts might have led to the responses19,20. Fourth, tumour hypoxia induced by anemia conflicting results. One study conducted in a single centre worsens oncologic outcomes by increasing proliferative and that used propensity score matching (n = 401) showed that metastatic potentials16. However, the theoretical detrimen- blood transfusion was not associated with worse oncologic tal effect of anemia has not always translated into poorer outcomes22. Another study based on population-based data oncologic outcomes in previous studies8–10,14,38. from a large number of cases (n = 24,330) showed that blood transfusion was related to worse outcomes43. Finally, not Perioperative Blood Transfusion all studies performed a multivariate analysis to control In crc, the perioperative transfusion rate has been reported for confounding factors39. We therefore reviewed earlier to be between 21.6% and 65.1%9,22–24,39–43; the rate in the studies that investigated the effect of preoperative anemia present study was 29.4%. The detrimental effect of blood and postoperative blood transfusion in nonmetastatic transfusion is thought to be attributable to transfusion- crc with multivariate analyses (details summarized in related immunomodulation and the systemic inflam- supplementary Table i). matory response. The immunologic changes induced by Tang et al.45 and Jagoditsch et al. 44 investigated prog- transfusion include decreased production of interleukin 2, nostic factors in crc by conducting multivariate analyses. inhibition of cytotoxic T cell activity, and increased immu- In both studies, preoperative anemia and perioperative nosuppressive prostaglandin release20. Those theoretical transfusion were used as variables in the univariate anal- disadvantages have not always translated into worse on- ysis; however, preoperative anemia was not included into cologic outcomes in previous studies. Indeed, the use of the multivariate analysis because it was nonsignificant in transfusion was related to worse os in several studies40–42, the univariate analysis. Moreover, those two studies failed with several other studies showing the opposite result9,24,44. to demonstrate any prognostic significance of preoperative

FIGURE 1 Overall survival by the Kaplan–Meier method. (A) The FIGURE 2 Recurrence-free survival by the Kaplan–Meier method. 3-year overall survival rate was significantly worse in the anemia group (A) The presence of preoperative anemia did not lead to a significantly than in the no-anemia group (81.5% vs. 87.5%, p = 0.002). (B) The worse recurrence-free survival rate (75.0% vs. 80.6%, p = 0.059). 3-year overall survival rate was worse for patients who received a (B) Receiving a perioperative allogenic red blood cell transfusion perioperative allogenic red blood cell transfusion than for those who did resulted in significantly worse recurrence-free survival (70.9% vs. not (76.9% vs. 88.2%, p < 0.001). (C) Taking preoperative anemia and 80.9%, p < 0.001). (C) The overall 3-year recurrence-free survival perioperative transfusion into consideration, the 3-year overall survival rates were 71.3%, 76.7%, 78.6%, and 82.4% in the no-anemia rates were 76.1%, 78.4%, 86.2%, and 89.4% in the anemia and trans- and transfusion, the anemia and transfusion, the anemia and no- fusion, the no-anemia and transfusion, the anemia and no-transfusion, transfusion, and the no-anemia and no-transfusion groups respec- and the no-anemia and no-transfusion groups respectively (p < 0.001). tively (p < 0.001).

Current Oncology, Vol. 26, No. 3, June 2019 © 2019 Multimed Inc. e363 ANEMIA AND TRANSFUSION IN COLORECTAL CANCER, Kwon et al. anemia or perioperative transfusion. In the study by Tang Morner et al.9 also used a multivariate analysis to et al.45, performed in 1993, the included patients had a investigate the effects of preoperative anemia and periop- mean age that was quite young compared with the age of erative transfusion. The purpose and design of their study the patients in other studies (no-transfusion group: 54.4 seem similar to ours, but their results are the opposite of ± 12.5 years; transfusion group: 55.2 ± 12.2 years), and ours. In their study, preoperative anemia was an indepen- only 14% had anemia, the lowest value in the reports that dent prognostic factor for os (hr: 2.3; p < 0.001) and rfs (hr: we reviewed. The presence of anemia and the detrimental 1.7; p < 0.05), but perioperative anemia was not. The Morner effect of transfusion might be less in young patients. The et al.9 study and our study have several differences. One is study by Jagoditsch et al.44, performed in 2006, included the number of patients. The Morner group acknowledge only patients with rectal cancer. In that study, preopera- the limited statistical power in their multivariate analy- tive anemia was defined as a hemoglobin level of 14 g/dL sis of risk factors for recurrence. The number of patients or less, the highest value in the reports we reviewed. The enrolled in their study was 496; our study enrolled 1003 percentage of patients with anemia was as high as 51.5%. patients. Another difference is the definition of the period Of all the reports we reviewed, the study by Talukder of “perioperative transfusion.” We defined perioperative et al.40, performed in 2014, included the largest number transfusion as receipt of a transfusion during postopera- of patients (n = 1370). In that study, 36.5% of the patients tive hospitalization. Hypothetically, transfusion induces presented with preoperative anemia (hemoglobin ≤ 12 g/ harmful immune reactions such as immunosuppression. dL), and 30.9% received a perioperative transfusion. The If transfusion causes an adverse immune reaction, that methods used by those authors had some distinct features. reaction could last until the postoperative period in affect- First, where we included 30-day postoperative morbidity in ed hosts. The healing process includes an inflammatory the multivariate analysis, they included variables related response in damaged tissue, secretion of growth factors, to the clinical circumstances requiring transfusion—for and cellular proliferation. Those reactions occur during the example, surgical urgency, medical complications, and postoperative period. We therefore believe that it is reason- reoperation. Second, they divided the perioperative period able to use a longer period (from the operative day to the into intraoperative, preoperative, and postoperative peri- date of discharge) for the use of transfusion. In the case of ods. They analyzed the effect on prognosis of transfusion the study by the Morner group, the period of perioperative during each period individually or during all periods in transfusion was defined as 24 hours before and after the combination. In their results, preoperative anemia was not operation. Third, despite the relatively short period defined a significant prognostic factor for os and rfs. Only intra- for transfusion, the transfusion rate was higher in the operative transfusion was significantly associated with os Morner group’s study than in our study (40.1% vs. 29.4%). (hr: 1.31; p = 0.004) and rfs (hr: 1.37; p = 0.013). Periopera- Fourth, their institution used leucocyte-depleted blood tive transfusion was not a significant prognostic factor for components, which are presumably less immunomodu- os, but it was significant for rfs (hr: 1.26; p = 0.024). They latory. Finally, the study locations were different (Europe explained that the loss of significance of perioperative vs. Asia). Those differences might have been responsible transfusion in the multivariate analysis for os could be for the conflicting results. attributed to confounding factors and the circumstances As we reviewed the published studies, we found that that necessitated transfusion rather than to the transfusion the effects of preoperative anemia and postoperative trans- itself. That analysis contradicts our results that periopera- fusion were reported differently. The inconsistent results tive transfusion is an independent prognostic factor for os between the studies might be attributable to heterogeneous regardless of unfavourable clinical confounding factors. study designs, patient cohorts, hemoglobin levels used to Their study is somewhat different from our study in that define preoperative anemia, and perioperative transfusion their study period was quite long (1984–2004), which sug- periods, in addition to complex treatment-related factors gests that patient treatments might heterogeneous. In fact, such as the surgical approach, the amount of intraoperative some patients received transfusions with whole blood and bleeding, and the indications for blood transfusion. not packed rbcs. Differences in the type of transfusion and We found that transfusion was negatively associated the definition of “perioperative transfusion” might have with oncologic outcomes. Thus, we believe that perioper- contributed to their different results. ative transfusion must be reduced. Well-organized treat- Kaneko et al.42 reported results similar to those in our ment processes must be recommended for patients who study about perioperative allogenic blood transfusion (hr: need a perioperative blood transfusion. 3.16; p = 0.031), but not preoperative hemoglobin, being an Even though preoperative anemia was not an inde- independent prognostic factor for os. However, their study pendent prognostic factor for survival in nonmetastatic included only a small number (n = 108) of elderly patients crc, it was a predictive factor for perioperative transfusion. (more than 75 years of age) and differed from other studies For patients with preoperative anemia, the indications for that analyzed preoperative hemoglobin rather than the blood transfusion and preoperative iron supplementa- presence of preoperative anemia. The variables used in tion should be standardized. During surgery, modifiable their multivariate analysis were tumour depth, lymph node treatment-related predictive factors that might mitigate the metastasis, preoperative hemoglobin, and perioperative need for blood transfusion—laparotomy, lengthy operative transfusion. Given that the study included only elderly time, and meticulous surgical techniques that minimize patients, we cannot compare their results with ours. In blood loss—must be considered. In addition, efforts must elderly patients, the adverse effect of blood transfusion be taken to reduce postoperative morbidity that requires might be greater. blood transfusion during the perioperative period.

Our study has several limitations. First, it has a retro- preoperative haemoglobin, C-reactive protein and albumin. spective, nonrandomized design. In addition, our database Colorectal Dis 2018;20:26–34. does not provide any information about the amount of 9. Morner ME, Edgren G, Martling A, Gunnarsson U, Egenvall M. intraoperative bleeding and details about blood trans- Preoperative anaemia and perioperative red blood cell transfusions given (such as the indications for transfusion, fusion as prognostic factors for recurrence and mortality in colorectal cancer—a Swedish cohort study. Int J Colorectal the volume of transfused blood, or other treatments for Dis 2017;32:223–32. anemia). Moreover, because our database does not include 10. McGrane JM, Humes DJ, Acheson AG, Minear F, Wheeler any information about the number of transfused units, we JMD, Walter CJ. Significance of anemia in outcomes after could not assess for any relationship between the transfu- neoadjuvant chemoradiotherapy for locally advanced rectal sion volume and oncologic outcomes. Although our study cancer. Clin Colorectal Cancer 2017;16:381–5. did not demonstrate a relationship between transfusion 11. An MS, Yoo JH, Kim KH, et al. T4 Stage and preoperative ane- volume and poor prognosis, it clearly demonstrated that mia as prognostic factors for the patients with colon cancer perioperative transfusion itself is associated with a detri- treated with adjuvant folfox chemotherapy. World J Surg mental effect on oncologic outcomes. Oncol 2015;13:64. 12. Zhen L, Zhe S, Zhenning W, et al. Iron-deficiency anemia: CONCLUSIONS a predictor of diminished disease-free survival of T3N0M0 stage colon cancer. J Surg Oncol 2012;105:371–5. 13. Stapley S, Peters TJ, Sharp D, Hamilton W. The mortality of The use of perioperative blood transfusion, but not preop- colorectal cancer in relation to the initial symptom at presen- erative anemia, was independently associated with worse tation to primary care and to the duration of symptoms: a co- os and rfs in nonmetastatic crc. Our findings indicate hort study using medical records. Br J Cancer 2006;95:1321–5. a need to reduce the use of blood transfusion during the 14. Berardi R, Braconi C, Mantello G, et al. Anemia may influence perioperative period for better oncologic outcomes. the outcome of patients undergoing neo-adjuvant treatment of rectal cancer. Ann Oncol 2006;17:1661–4. ACKNOWLEDGMENTS 15. Tampellini M, Saini A, Alabiso I, et al. The role of haemoglobin This research was supported by the Basic Science Research Pro- level in predicting the response to first-line chemotherapy in gram through the National Research Foundation of Korea (nrf) advanced colorectal cancer patients. Br J Cancer 2006;95:13–20. funded by the Ministry of Education (nrf 2017R1D1A3B03032301). 16. Khan AA, Klonizakis M, Shabaan A, Glynne-Jones R. As- sociation between pretreatment haemoglobin levels and CONFLICT OF INTEREST DISCLOSURES morphometric characteristics of the tumour, response to We have read and understood Current Oncology’s policy on dis- neoadjuvant treatment and long-term outcomes in pa- closing conflicts of interest, and we declare that we have none. tients with locally advanced rectal cancers. Colorectal Dis 2013;15:1232–7. AUTHOR AFFILIATIONS 17. Fjortoft I, Furnes B, Hausken T, Storli KE, Eide GE, Sondenaa *Division of Colorectal Surgery, Department of Surgery, Yonsei K. Pre-operative anaemia in colon cancer patients became University Wonju College of Medicine, †Health Promotion Center, normal after more than a year post-operatively but did not Wonju Severance Christian Hospital, ‡Division of Gastroenterol- influence oncological outcome in the final analysis. Scand J ogy, Department of Internal Medicine, Yonsei University Wonju Gastroenterol 2013;48:663–71. College of Medicine, and §Big Data Research Group, Yonsei Uni- 18. Roxburgh CS, Wallace AM, Guthrie GK, Horgan PG, McMil- versity Wonju College of Medicine, Wonju, R.O.K. lan DC. Comparison of the prognostic value of tumour- and patient-related factors in patients undergoing potentially cu- REFERENCES rative surgery for colon cancer. Colorectal Dis 2010;12:987–94. 1. Ferlay J, Soerjomataram I, Dikshit R, et al. Cancer incidence 19. Opelz G, Sengar DP, Mickey MR, Terasaki PI. Effect of blood and mortality worldwide: sources, methods and major transfusions on subsequent kidney transplants. Transplant patterns in globocan 2012. Int J Cancer 2015;136:E359–86. Proc 1973;5:253–9. 2. Wan Kim Y. Surgical treatment for colorectal cancer in octo- 20. Cata JP, Wang H, Gottumukkala V, Reuben J, Sessler DI. genarians and nonagenarians. J BUON 2017;22:578–85. Inflammatory response, immunosuppression, and cancer 3. Wilson MJ, van Haaren M, Harlaar JJ, et al. Long-term recurrence after perioperative blood transfusions. Br J An- prognostic value of preoperative anemia in patients with aesth 2013;110:690–701. colorectal cancer: a systematic review and meta-analysis. 21. Amato A, Pescatori M. Perioperative blood transfusions for Surg Oncol 2017;26:96–104. the recurrence of colorectal cancer. Cochrane Database Syst 4. World Health Organization (who). Haemoglobin Concentra- Rev 2006;:CD005033. tions for the Diagnosis of Anaemia and Assessment of Severity. 22. Warschkow R, Guller U, Koberle D, et al. Perioperative blood Geneva, Switzerland: who; 2011. [Available online at: http:// transfusions do not impact overall and disease-free survival www.who.int/vmnis/indicators/haemoglobin.pdf; cited after curative rectal cancer resection: a propensity score 10 January 2019] analysis. Ann Surg 2014;259:131–8. 5. Sadahiro S, Suzuki T, Tokunaga N, et al. Anemia in patients 23. Tarantino I, Ukegjini K, Warschkow R, et al. Blood transfusion with colorectal cancer. J Gastroenterol 1998;33:488–94. does not adversely affect survival after elective colon cancer 6. Weiss G, Goodnough LT. Anemia of chronic disease. N Engl resection: a propensity score analysis. Langenbecks Arch Surg J Med 2005;352:1011–23. 2013;398:841–9. 7. Vayrynen JP, Tuomisto A, Vayrynen SA, et al. Preoperative 24. Ghinea R, Greenberg R, White I, Sacham-Shmueli E, Mahagna anemia in colorectal cancer: relationships with tumor char- H, Avital S. Perioperative blood transfusion in cancer patients acteristics, systemic inflammation, and survival. Sci Rep undergoing laparoscopic colorectal resection: risk factors 2018;8:1126. and impact on survival. Tech Coloproctol 2013;17:549–54. 8. Egenvall M, Morner M, Martling A, Gunnarsson U. Predic- 25. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, tion of outcome after curative surgery for colorectal cancer: Vandenbroucke JP on behalf of the strobe Initiative. The

Strengthening the Reporting of Observational Studies in 36. Munoz M, Gomez-Ramirez S, Martin-Montanez E, Auerbach Epidemiology (strobe) statement: guidelines for reporting M. Perioperative anemia management in colorectal can- observational studies. Lancet 2007;370:1453–7. cer patients: a pragmatic approach. World J Gastroenterol 26. Dindo D, Demartines N, Clavien PA. Classification of sur- 2014;20:1972–85. gical complications: a new proposal with evaluation in a 37. Pahor M, Manini T, Cesari M. Sarcopenia: clinical evaluation, cohort of 6336 patients and results of a survey. Ann Surg biological markers and other evaluation tools. J Nutr Health 2004;240:205–13. Aging 2009;13:724–8. 27. Kim IY, Kim BR, Kim YW. Impact of prior abdominal surgery 38. van Halteren HK, Houterman S, Verheij CD, Lemmens VE, on rates of conversion to open surgery and short-term out- Coebergh JW. Anaemia prior to operation is related with comes after laparoscopic surgery for colorectal cancer. PLoS poorer long-term survival in patients with operable rectal One 2015;10:e0134058. cancer. Eur J Surg Oncol 2004;30:628–32. 28. Kim IY, You SH, Kim YW. Neutrophil–lymphocyte ratio pre- 39. Gunka I, Dostalik J, Martinek L, Gunkova P, Mazur M. Impact dicts pathologic tumor response and survival after preoper- of blood transfusions on survival and recurrence in colorectal ative chemoradiation for rectal cancer. BMC Surg 2014;14:94. cancer surgery. Indian J Surg 2013;75:94–101. 29. Kim DH, Kim IY, Kim BR, Kim YW. Factors affecting the se- 40. Talukder Y, Stillwell AP, Siu SK, Ho YH. Comparing survival lection of minimally invasive surgery for stage 0/I colorectal and recurrence in curative stage i to iii colorectal cancer in cancer. Int J Surg 2015;16:44–8. transfused and nontransfused patients. Int Surg 2014;99:8–16. 30. Kim IY, Kim BR, Kim YW. The short-term and oncologic 41. Li XX, Meng J, Sun GP, et al. Effects of perioperative blood outcomes of laparoscopic versus open surgery for T4 colon transfusion on the prognosis in hereditary and sporadic cancer. Surg Endosc 2016;30:1508–18. colon cancer. Biomarkers 2015;20:481–6. 31. Kim IY, Kim BR, Choi EH, Kim YW. Short-term and oncologic 42. Kaneko M, Sasaki S, Ishimaru K, Terai E, Nakayama H, outcomes of laparoscopic and open complete mesocolic Watanabe T. The impact of perioperative allogeneic blood excision and central ligation. Int J Surg 2016;27:151–7. transfusion on survival in elderly patients with colorectal 32. National Comprehensive Cancer Network (nccn). NCCN cancer. Anticancer Res 2015;35:3553–8. Clinical Practice Guidelines in Oncology: Rectal Cancer. 43. Aquina CT, Blumberg N, Becerra AZ, et al. Association among Ver. 1.2018. Fort Washington, PA: nccn; 2018. [Current version blood transfusion, sepsis, and decreased long-term survival available online at: http://www.nccn.org/professionals/ after colon cancer resection. Ann Surg 2017;266:311–17. physician_gls/pdf/rectal.pdf (free registration required); 44. Jagoditsch M, Pozgainer P, Klingler A, Tschmelitsch J. Impact cited 10 January 2019] of blood transfusions on recurrence and survival after rectal 33. National Comprehensive Cancer Network (nccn). NCCN cancer surgery. Dis Colon Rectum 2006;49:1116–30. Clinical Practice Guidelines in Oncology: Colon Cancer. 45. Tang R, Wang JY, Chien CR, Chen JS, Lin SE, Fan HA. The asso- Ver. 2.2018. Fort Washington, PA: nccn; 2018. [Current version ciation between perioperative blood transfusion and survival available online at: http://www.nccn.org/professionals/ of patients with colorectal cancer. Cancer 1993;72:341–8. physician_gls/pdf/colon.pdf (free registration required); 46. Tartter PI. The association of perioperative blood transfusion cited 10 January 2019] with colorectal cancer recurrence. Ann Surg 1992;216:633–8. 34. Makipour S, Kanapuru B, Ershler WB. Unexplained anemia 47. Miki C, Hiro J, Ojima E, Inoue Y, Mohri Y, Kusunoki M. Periop- in the elderly. Semin Hematol 2008;45:250–4. erative allogeneic blood transfusion, the related cytokine 35. Ho CH, Yu YB, Wu PH. The prevalence of iron deficiency ane- response and long-term survival after potentially curative mia and its clinical implications in patients with colorectal resection of colorectal cancer. Clin Oncol (R Coll Radiol) carcinoma. J Chin Med Assoc 2008;71:119–22. 2006;18:60–6.